An energy efficient and high speed architecture for convolution computing based on binary resistive random access memory
Authors:
- Chen LiuCLChen Liu
- Runze HanRHRunze Han
- Zheng ZhouZZZheng Zhou
- Peng HuangPHPeng Huang
- Lifeng LiuLLLifeng Liu
- Xiaoyan LiuXLXiaoyan Liu
- Jinfeng KangJKJinfeng Kang
- 6 March 2018
- journal article
- conference paper
- Published by IOP Publishing in Japanese Journal of Applied Physics
- Vol. 57 (4S) , 04FE05
Abstract
In this work we present a novel convolution computing architecture based on metal oxide resistive random access memory (RRAM) to process the image data stored in the RRAM arrays. The proposed image storage architecture shows performances of better speed-device consumption efficiency compared with the previous kernel storage architecture. Further we improve the architecture for a high accuracy and low power computing by utilizing the binary storage and the series resistor. For a 28 × 28 image and 10 kernels with a size of 3 × 3, compared with the previous kernel storage approach, the newly proposed architecture shows excellent performances including: 1) almost 100% accuracy within 20% LRS variation and 90% HRS variation; 2) more than 67 times speed boost; 3) 71.4% energy saving.Keywords
Funding Information
- National Natural Science Foundation of China (61334007, 61604005, 61421005)
This publication has 31 references indexed in Scilit:
- Synaptic electronics: materials, devices and applicationsNanotechnology, 2013
- The parallel approachNature Physics, 2013
- HfOx-Based Vertical Resistive Switching Random Access Memory Suitable for Bit-Cost-Effective Three-Dimensional Cross-Point ArchitectureACS Nano, 2013
- A Low Energy Oxide‐Based Electronic Synaptic Device for Neuromorphic Visual Systems with Tolerance to Device VariationAdvanced Materials, 2013
- Stochastic learning in oxide binary synaptic device for neuromorphic computingFrontiers in Neuroscience, 2013
- The future of microprocessorsCommunications of the ACM, 2011
- Nanoionics-based resistive switching memoriesNature Materials, 2007
- Magnetic Domain-Wall LogicScience, 2005
- Programmable computing with a single magnetoresistive elementNature, 2003
- Can programming be liberated from the von Neumann style?Communications of the ACM, 1978